402 ALASKAN GLACIER STUDIES more actively into the river bank than they had done in previous years. A river bank about 25 feet high was cut back for 20 feet in one place, though elsewhere the average retreat of the gravel bluff was not more than 3% to 5 feet during the summer. When the ice front of the advancing glacier was sufficiently undercut by the river the ice slid down into the water with a splash, forming a wave which rushed with great force across the river and up on the opposite bank. In 1908 one such wave deposited a bowlder weighing approximately 200 pounds in the forks of a tree 80 feet back from the edge of the bank. Owing to the decreased width of the river in 1909, however, the waves dashed back over the bank for a much greater distance than in 1908, tearing away earth, bowlders, and trees, and breaking or knocking the bark from trees 60 feet from the river. That the 1909 rate of motion was slow compared with that a year later is indicated by the fact that the engineers were able to go out on the glacier every few days in July to reset the flags for measuring the rate of motion, while in 1910 this portion of the ice was so severely crevassed that this could not be done. From these facts Mr. Johnson concluded that the glacier was pushing forward and forcing the river eastward in the summer of 1909; and this commencement of advance was announced by us after our visit to Childs Glacier in 1909 .* Spasmodic Advance in 1910. The rate of motion in Childs Glacier increased during the winter of 1909-1910 so that part of the margin, hitherto stagnant, attained a forward movement of from two to eight feet a day, while in mid-glacier the rate of movement increased at least five or six times.8 The ice which covers Copper River in winter serves as a delicate index of the rate of motion of Childs Glacier during that season. The railway engineers testify that in winter the river ice has always been broken more or less by advance of the glacier, and the description by Allen in 1885 and a photograph by McPherson in January, 1906, shows that there was slight breakage of the river ice at these times. During the winter and early spring of 1909-10, the glacier moved forward rapidly enough to buckle up the ice of the frozen river much more than usual, which is interpreted as proof that the advance of Childs Glacier continued throughout that season. By June 10, 1910, the ice front in various portions near mid-glacier had moved forward from 275 to 650 feet since the preceding autumn; but the total advance, including that of 1909, was about 1200 feet. Between the measurements of July, 1909, and June 3, 1910, the river had been narrowed down to a width of from 500 to 650 feet, and, if by this constriction the eastern river bank had been cut back, the amount of advance stated above is to be considered a minimum to which it may be necessary to add from 40 to 60 feet. As a result of the constriction resulting from this advance the water rose 21 feet in a few days. People locally speak of the discharge from the front of Childs Glacier as "sloughing." A "slough" has always raised waves in Copper River (PI. CXLIX), making it dangerous for a boat to shoot the rapids in front of the glacier, or to line a boat up the opposite bank; but in the spring of 1910 the dangerous conditions were accentuated by the advance of the glacier and the pushing of the river eastward. In that year the waves i Martin, Lawrence, in H. F. Reid's Variations of Glaciers for 1909, Journ. Geol., Vol. X.IX, 1911, p. 88; Zeitschrift fidr Gletacherkunde, Band V, 1911, pp. 200-201. »Martin. Lawrence, Nat. Geog. Mag., Vol. XXH, 1911, pp. 541-548; in H. F. Reid's Variations of Glaciers for 1910, Journ. Geog., Vol. XIX, 1911, pp. 458-459; Zeitschrift fur Gletscherkunde, Band VI, 1911, pp. 102-103.